search

slarson / wholebrain

Automatically exported from code.google.com/p/wholebrain
0 stars 0 forks source link

Cell networks not coming up on production #491

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
    Hi Dave,

       It looks like we may have a problem with production.  The cell files
that we see coming up on the stage environment are not there on prod.  To test:

       1. Bring up production client 0.7.7.0
       2. In the data window, open the tree called "cells"
       3. Click on the check box next to "Granule cell layer"
       4. Open up the tree called "granule cell layer"
       5. Double click on any of the items underneath.

    You should see cells under there, but instead you don't get any. 
Perhaps the logs will reveal what is going on?  Missing data files?

    Thanks,
      Stephen

Original issue reported on code.google.com by stephen....@gmail.com on 11 May 2010 at 7:21

GoogleCodeExporter commented 9 years ago 
Hi Dave & all

     OK, to fix this problem for the time being, I am going to hard code the main
production client to hit against stage-data.  To see this behavior, please use 
the
latest dev client and manually point it at data.wholebraincatalog.org

Thanks,
  Stephen

Original comment by stephen....@gmail.com on 11 May 2010 at 7:22

GoogleCodeExporter commented 9 years ago 

[12:05] <slars0n_> hey guys; can you guys load the production client
[12:05] <caprea> ill give it a shot
[12:05] <slars0n_> and then click on (make visible) a cell network?  like 
granule
cell layer
[12:05] <slars0n_> and then try to zoom to a cell or find the network?
[12:07] <jrmartin> i tried it
[12:07] <jrmartin> can't zoom in to cells
[12:07] <caprea> it didnt go for granule cell
[12:07] <jrmartin> in fact, cells aren't coming up
[12:07] <slars0n_> ok; looking at stage now
[12:08] <slars0n_> let's figure out where the issue happened
[12:08] <caprea> yeah no cells, gonig to look at the log file
[12:08] <slars0n_> looks like stage is working
[12:08] <slars0n_> weird
[12:08] <slars0n_> perhaps there was a problem migrating from stage to prod
[12:08] <slars0n_> can you guys verify it works on stage?
[12:08] <jrmartin> yeah seems like server problem to me
[12:08] <jrmartin> i am trying it right now
[12:10] <jrmartin> stage works well for me
[12:10] <caprea> cells turn on and off, Jesus?
[12:10] <jrmartin> yes
[12:10] <caprea> okay, still investigating on logfile
[12:11] <slars0n_> what log are you looking at?
[12:11] <caprea> client log
[12:12] <slars0n_> ah ok
[12:12] <slars0n_> I'm sending dave an e-mail now.  please attach anything you 
find
to it.
[12:12] <slars0n_> can one of you see if there is anything else missing in prod?
[12:13] <slars0n_> i'm looking around on it but can't find anything else
[12:13] <caprea> im going that now
[12:13] <caprea> im lookign at datawrappers and such
[12:14] <caprea> everything looks okay so far, going to run client in console 
to see
more info
[12:16] <slars0n_> ok
[12:16] <slars0n_> i'm thinking i'm going to point the prod client at stage for 
the
time being.  objections?
[12:18] <caprea> the two cells that are not in a network come up
[12:18] <slars0n_> weird
[12:18] <jrmartin> not for me
[12:19] <caprea> I just did a fresh run, toggled them first
[12:20] <jrmartin> i see, they do come up only when selecting on them 
individually
[12:20] <jrmartin> dentate gyrus network comes up
[12:21] <caprea> I didnt get it to come up
[12:21] <caprea> will try again

Original comment by stephen....@gmail.com on 11 May 2010 at 7:22

GoogleCodeExporter commented 9 years ago 
May 11, 2010 12:23:29 PM java.util.logging.LogManager$RootLogger log
WARNING: Estimated filesize is: 6KB (6867 B)
May 11, 2010 12:23:29 PM java.util.logging.LogManager$RootLogger log
WARNING: Estimated filesize is: 6KB (6303 B)
May 11, 2010 12:23:29 PM java.util.logging.LogManager$RootLogger log
WARNING: Estimated filesize is: 2KB (2788 B)
java.lang.ClassCastException: java.io.File cannot be cast to
org.morphml.neuroml.schema.Level3Cell
    at
org.wholebrainproject.wbc.tangible.neuronmorphology.NeuronMorphology.getData(Neu
ronM
orphology.java:522)
    at
org.wholebrainproject.wbc.tangible.neuronmorphology.NeuronMorphology.initializeC
able
IdToSegmentIdMap(NeuronMorphology.java:607)
    at
org.wholebrainproject.wbc.tangible.neuronmorphology.NeuronMorphology.getCableIdT
oSeg
mentIdMap(NeuronMorphology.java:551)
    at
org.wholebrainproject.wbc.tangible.neuronmorphology.NeuronMorphology.getPartCoun
t(Ne
uronMorphology.java:597)
    at
org.wholebrainproject.wbc.view.tangible.NeuronCloud.makeHDInstances(NeuronCloud.
java
:409)
    at
org.wholebrainproject.wbc.view.tangible.NeuronCloud.setup(NeuronCloud.java:640)
    at
org.wholebrainproject.wbc.view.tangible.NeuronCloud.addCloudInstances(NeuronClou
d.ja
va:826)
    at org.wholebrainproject.wbc.view.View3D.setCells(View3D.java:857)
    at
org.wholebrainproject.wbc.observers.SceneViewObserver.reloadCells(SceneViewObser
ver.
java:428)
    at
org.wholebrainproject.wbc.observers.SceneViewObserver.buildOutstanding(SceneView
Obse
rver.java:395)
    at
org.wholebrainproject.wbc.view.View3D$TangibleViewMaker.run(View3D.java:199)
    at java.lang.Thread.run(Thread.java:619)
May 11, 2010 12:23:30 PM java.util.logging.LogManager$RootLogger log
WARNING: Finished creating meshes for the neuron cloud 19 ms.

Original comment by caprea on 11 May 2010 at 7:24

GoogleCodeExporter commented 9 years ago 
We've just realized that the exception posted above is not related to the 
problem; its  
a red herring.  Please disregard

Original comment by stephen....@gmail.com on 11 May 2010 at 11:25

GoogleCodeExporter commented 9 years ago 
The exception was somewhat related - I made a code change and it allowed the 
cells
to appear on data, and data2. It worked on both servers.

The code change I made was to the client-side caching, I was accidentally 
caching
the file LOCATION instead of the content itself.

BUT, this is not a satisfactory bug fix. because it was working on stage, dev 
and
data2. I can't explain why it should work with some servers and not others. 
Although
there is a fix for client-side, this issue is still unresolved.

Original comment by caprea on 12 May 2010 at 12:26

GoogleCodeExporter commented 9 years ago 
Summary:

- Team reported problem with prod
- stage and dev worked correctly
- Larry rebooted data, problem persisted on data
- Dave reinstalled data2 by pushing from stage to data2, problem persisted on 
data
- Dave did not reinstall data1
- Jesus again tested data2 and data1 (bypassing data)
- When data (the LB) was not involved (i.e. dev, stage, data2 (bypass data), 
and 
data1 (bypass data)), everything worked correctly
- When data was involved (i.e. pass through to data2) the problem appeared.
- Vicky and Dave can demo correct and incorrect behavior
- Server logs on all server show no evidence of errors

Dave.

Original comment by stephen....@gmail.com on 12 May 2010 at 2:33

GoogleCodeExporter commented 9 years ago

Original comment by stephen....@gmail.com on 12 May 2010 at 2:33

GoogleCodeExporter commented 9 years ago 
our fix was to use the IP address to reference production.  we expect that with 
client 
fixes this will not be a problem going forward

Original comment by stephen....@gmail.com on 13 May 2010 at 10:39